基于离散HMM的非特定人关键词提取语音识别系统

基于离散HMM的非特定人关键词提取语音识别系统

汪鹏，刘加，刘润生

清华大学电子工程系, 北京 100084

收稿日期:2002-12-04 修回日期:1900-01-01 出版日期:2003-07-26 发布日期:2003-07-26

Discrete HMM Based Speaker Independent Keyword Spotting Speech Recognition Syste m

WANG Peng, LIU Jia, LIU Run-sheng

Department of Electronic Engineering, Tsinghua University, Beijing 100084, China

Received:2002-12-04 Revised:1900-01-01 Online:2003-07-26 Published:2003-07-26

摘要/Abstract

摘要： 设计了一个基于离散隐含马尔可夫整词模型的非特定人汉语小词表关键词提取语音识别系统, 提出一种基于对前向、后向搜索得到的关键词假设做二次识别的置信度策略. 将该置信度用于关键词提取系统的说话验证, 得到了良好效果.

关键词: 语音识别, 关键词提取, 隐含马尔可夫模型, 置信度, 拒识

Abstract: This paper describes a system of small-vocabulary Mand arin keyword spotting speech recognition based on discrete HMMs. A novel confide nce measure for utterance verification is also presented. The confidence score i s obtained by rescoring putative keywords detected by the keyword spotter via a Viterbi search algorithm on the isolated keyword HMMs. Experimental results show that the confidence measure has good performance on discriminating keyword detections from false alarms.

Key words: speech recognition, keyword spotting, HMM, confidence score, rejection

中图分类号:

TN912.34

汪鹏，刘加，刘润生. 基于离散HMM的非特定人关键词提取语音识别系统[J]. J4, 2003, 41(03): 347-351.

WANG Peng, LIU Jia, LIU Run-sheng. Discrete HMM Based Speaker Independent Keyword Spotting Speech Recognition Syste m[J]. J4, 2003, 41(03): 347-351.

[1]	董立岩, 朱琪, 隋鹏, 孙鹏, 李永丽. 基于最大差距的置信度评估算法[J]. 吉林大学学报(理学版), 2015, 53(06): 1217-1222.
[2]	刘妍秀, 孙一鸣, 杨华民. 基于归一化算法的噪音鲁棒性连续语音识别[J]. 吉林大学学报(理学版), 2015, 53(03): 519-524.
[3]	崔明, 许志闻. 基于加权DFFD算法和渐变动画思想的人脸动画系统[J]. J4, 2012, 50(02): 288-292.
[4]	吴玺宏, 吴昊, 高勤, 林小俊, 王馨浩. 潜语义分析语言模型及其在汉语大词汇连续语音识别中的应用[J]. J4, 2006, 44(06): 16-20.